Any interstellar method is a fortiori a terrestrial method as well.
It's a task with entirely different challenges and to which entirely different tactics are optimal. It is just plain not the case that all interstellar methods are a fortiori terrestrial methods - since the very method your declaration of impossibility assumes is sufficiently optimal to rule out any possibility is one that doesn't work when launched from the surface of the target planet.
The Tsar Bomba gives us an estimate on what the third state buys one, let's be generous roughly an order of magnitude (6mt to 60mt).
The Tsar Bomba was deliberately crippled from it's original specifications so as to reduce fallout while still being a sufficiently excessive explosion (although this was only on the order of a twofold reduction). Regardless I am almost as willing to accept the Tsar Bomba as the theoretical best (or worst) case for what humankind can do with a three stage thermonuclear device as I am to declare that 'Little Boy' the most potent fissile device that humans could make.
Simply taking the calculations used for hypothesizing about Orion weapons and applying them to surface based doomsday device is a gross error. The approach taken and the limiting factors would be entirely different.
This is not to say that I believe we currently have the technology to destroy the planet itself. To the best of my knowledge we do not. I simply reject the notions that an Orion device would be the way to go about it and hence that the fission limitation found when hypothesizing about Orion still applies at the levels calculated. I don't present an alternate hypothesis that because these limitations no longer apply as calculated that we must be able to do the destruction - that would be reversing stupidity.
And where does the Orion vehicle hit?
Not at the location where the bombs go off - that's exactly the problem.
(Would an ICBM be excluded as an answer because it enters space?)
I suppose Multi would allow it provided in returns and detonates at or near enough to the surface. My initial speculations about using Orion for self destruction speculated at finding a gravity well sufficient to slingshot it back to us - but that engineering feat strikes me as a tad difficult.
It is just plain not the case that all interstellar methods are a fortiori terrestrial methods
I'll rephrase it, then: any method planet A can use to attack planet B can also be used by planet A to attack planet A. In the case of an Orion object, some large hyperbolic orbit which intersects the Earth seems possible, but if not, it could always be slowly launched out to whatever distance necessary and turn around.
...The Tsar Bomba was deliberately crippled from it's original specifications so as to reduce fallout while still being a sufficiently excessive
http://neo.jpl.nasa.gov/risk/2012da14.html
http://rt.com/news/paint-asteroid-earth-nasa-767/
Seems like a good opportunity to bring up existential risks. And A friendly reminder that NASA is in fact pretty damned important.
Thoughts?